Bivariate Poisson Model Using Multi-Level Method and Its Application in Analyzing the Factors Affecting Sick Leave of Employees of a Steel Company

Background: One of the issues that affects the economic productivity is the absence from work due to illness (sick leave). Considering the importance of this subject, in this study, the factors affecting the number of days of sick leave due to occupational and internal diseases have been investigated. Methods: This cross-sectional study was performed on male employees working in Mobarakeh Steel Company in Isfahan, Iran, from 2011 to 2015. In this study, the response variables were the number of days of sick leave due to internal and occupational diseases, and the covariate variables were shift work, smoking status, education, age, work experience, body mass index (BMI), and Framingham Risk Score (FRS). Data were analyzed by bivariate Poisson model using multi-level method. Findings: The present study was performed on 17988 male workers with mean age of 38.13 ± 7.56 years and mean work experience of 6.79 ± 5.95 years. The results showed that the variables of shift work (eβ = 1.47), smoking (eβ = 1.82), education (eβ = 1.11), age (eβ = 1.02), work experience (eβ = 1.05), BMI (eβ = 1.02), and FRS (eβ = 1.08) had a significant effect on the average sick leave due to internal diseases; however, of the above variables, only four variables of smoking (eβ = 0.74), education (eβ = 1.19), work experience (eβ = 1.04), and FRS (eβ = 1.02) had a significant effect on the average sick leave due to occupational diseases. Conclusion: According to the results of this study, by controlling the effective variables, it is possible to provide appropriate strategies to reduce the number of days of sick leave.